螺旋断层放射治疗系统兆伏级CT图像的增强优化

Image Enhancement of Megavolt CT in Tomotherapy System

目的针对螺旋断层放射治疗系统采集的兆伏级CT(MVCT)图像质量差,导致自适应放射治疗(DGART)时无法精确定位肿瘤靶区的问题,提出应用计算机图像处理技术对MVCT图像作增强优化处理,并验证其可行性。方法选择2017年2月至2019年12月于医院采用螺旋断层放射治疗(Tomo)系统行放射治疗的10例乳腺癌保乳术后患者作为研究对象,调取所有患者的首日摆位MVCT图像,联合应用非局域均值滤波算法和反锐化掩模算法对其作去噪和锐化图像增强处理,比较未处理的MVCT图像、增强MVCT图像及CT模拟机定位时的千伏级CT(kVCT)图像的清晰度和器官轮廓识别度。结果与未处理的MVCT图像比较,增强MVCT图像的清晰度得到了较大提高,且图像中组织器官的视觉效果接近于定位kVCT图像。在增强MVCT图像上勾画的乳腺癌靶区与定位kVCT图像上勾画的靶区的形状相似性指数(DSC)高于0.93。结论利用非局域均值滤波算法和反锐化掩模算法对Tomo系统的MVCT图像进行增强优化是可行的,增强MVCT图像的质量可达到临床乳腺癌保乳术后放射治疗靶区勾画的标准。

Objective Considering that the delineation accuracy would be affected by the poor quality of the original MVCT images in the subsequent adaptive radiotherapy,the method of applying computer image processing technology to enhance megavolt CT(MVCT)images in tomotherapy system was proposed and its feasibility was verified.Methods Totally 10 breast cancer patients who were treated with tomotherapy(Tomo)after breast-conserving surgery were chosen as the subjects in this study.Their first-day MVCT images were collected and then non-local mean filtering algorithm combined with unsharp masking algorithm were used to enhance the MVCT images by denoising and sharpening.Meanwhile,the image clarity and recognition of the organs’outlines were compared among the original MVCT images,the enhanced MVCT images and the kilovolt CT(kVCT)images acquired during the positioning by the CT analogue machine.Results Compared with the original MVCT images,the clarity of the enhanced MVCT images has greatly been improved,and the visual effect of the organs in the enhanced MVCT images were close to the positioning kVCT images.The similarity coefficient of the delineated target volumes between the enhanced MVCT and the kVCT was higher than 0.93.Conclusion It is feasible to use non-local mean filtering and unsharp masking algorithms to enhance and optimize the MVCT images,and the image quality could reach the standard of delineating the target area of radiotherapy after breast-conserving surgery.